fix

cmd/server/main.go

@@ -244,6 +244,16 @@ func main() {
 	historySize := 100
 	store := setupStore(context.Background())
 	coll := collector.New(nil, historySize, store) // Temporary nil for wsServer
+	if occurrenceStore, ok := store.(storage.OccurrenceStore); ok {
+		occurrences, err := occurrenceStore.RecentOccurrences(context.Background(), 100000)
+		if err != nil {
+			log.Printf("issues: persisted index restore skipped: %v", err)
+		} else if n, err := coll.RestoreIssues(context.Background(), occurrences); err != nil {
+			log.Printf("issues: persisted index restore failed after %d occurrences: %v", n, err)
+		} else if len(occurrences) > 0 {
+			log.Printf("issues: restored %d persisted occurrences", n)
+		}
+	}
 	kubeWatcher, kubeErr := kube.NewFromEnv(context.Background())
 	if kubeErr != nil {
 		log.Printf("kubernetes: rollout watcher disabled: %v", kubeErr)

internal/alerts/notify.go

@@ -312,7 +312,23 @@ func (a Alert) HTML(baseURL string, max int) string {
 			fmt.Fprintf(&s, "🔗 <a href=\"%s\">Open issue</a>\n", esc(base+"/issues/"+url.PathEscape(p.IssueID)))
 		}
 		if plainLen(b.String())+plainLen(s.String())+plainLen(foot)+40 > max {
-			fmt.Fprintf(&b, "\n… and %d more pattern%s, see the issues list", len(a.Patterns)-i, plural(len(a.Patterns)-i))
+			remaining := len(a.Patterns) - i
+			b.WriteString("\n<b>Other detected patterns</b>\n")
+			for _, other := range a.Patterns[i:] {
+				var compact strings.Builder
+				fmt.Fprintf(&compact, "· <code>%s</code> — %d now, normally %.0f", esc(other.What), other.Count, other.Baseline)
+				if base != "" && other.IssueID != "" {
+					fmt.Fprintf(&compact, " · <a href=\"%s\">open issue</a>", esc(base+"/issues/"+url.PathEscape(other.IssueID)))
+				}
+				compact.WriteByte('\n')
+				if plainLen(b.String())+plainLen(compact.String())+plainLen(foot)+10 > max {
+					break
+				}
+				b.WriteString(compact.String())
+			}
+			if remaining > 0 && !strings.Contains(b.String(), "open issue") && base != "" {
+				fmt.Fprintf(&b, "(%d additional pattern%s; open Issues for the full list)", remaining, plural(remaining))
+			}
 			break
 		}
 		b.WriteString(s.String())

internal/collector/collector.go

@@ -6,6 +6,7 @@ import (
 	"container/ring"
 	"context"
 	"encoding/json"
+	"fmt"
 	"log"
 	"os"
 	"strconv"
@@ -230,6 +231,22 @@ func (c *Collector) SetIssueStatusStore(ctx context.Context, store issues.Status
 	return c.issues.SetStatusStore(ctx, store)
 }
 
+// RestoreIssues rebuilds the bounded in-memory issue index from persisted
+// occurrences. This keeps /issues aligned with alerting after a restart.
+func (c *Collector) RestoreIssues(ctx context.Context, occurrences []storage.Occurrence) (int, error) {
+	for _, occurrence := range occurrences {
+		var sample issues.Sample
+		if err := json.Unmarshal(occurrence.Sample, &sample); err != nil {
+			return 0, fmt.Errorf("decode persisted issue sample: %w", err)
+		}
+		if sample.Message == "" {
+			continue
+		}
+		c.issues.Restore(occurrence.Kind, occurrence.Service, sample.Message, sample)
+	}
+	return len(occurrences), nil
+}
+
 func (c *Collector) SimilarIssues(id string, limit int) ([]issues.SimilarIssue, bool) {
 	return c.issues.Similar(id, limit)
 }

internal/issues/tracker.go

@@ -227,6 +227,14 @@ func (t *Tracker) Record(kind, service, title string, sample Sample) {
 	}
 }
 
+// Restore records an occurrence loaded from the persistent occurrence store.
+// It deliberately does not invoke the occurrence sink, avoiding duplicate
+// writes while rebuilding the in-memory index during startup.
+func (t *Tracker) Restore(kind, service, title string, sample Sample) {
+	normalizeSample(&sample, title)
+	_, _ = t.record(kind, service, title, sample)
+}
+
 func (t *Tracker) record(kind, service, title string, sample Sample) (string, OccurrenceSink) {
 	id := Fingerprint(kind, service, NormalizeTitle(title))
 	bucket := HourBucket(sample.Timestamp)

internal/storage/clickhouse_traces.go

@@ -213,6 +213,44 @@ func (s *ClickHouseStore) AppendOccurrence(ctx context.Context, occ Occurrence)
 	return nil
 }
 
+// RecentOccurrences returns persisted issue occurrences in chronological
+// order so the in-memory issue tracker can be rebuilt after a restart.
+func (s *ClickHouseStore) RecentOccurrences(ctx context.Context, limit int) ([]Occurrence, error) {
+	if limit <= 0 {
+		limit = 100000
+	}
+	rows, err := s.conn.Query(ctx, fmt.Sprintf(`
+		SELECT issue_id, kind, service, toUnixTimestamp64Nano(timestamp), trace_id, span_id, sample
+		FROM %s.issue_occurrences
+		ORDER BY timestamp DESC
+		LIMIT ?`, s.db), limit)
+	if err != nil {
+		return nil, fmt.Errorf("clickhouse recent issue occurrences: %w", err)
+	}
+	defer rows.Close()
+	var out []Occurrence
+	for rows.Next() {
+		var o Occurrence
+		var ts int64
+		var sample string
+		if err := rows.Scan(&o.IssueID, &o.Kind, &o.Service, &ts, &o.TraceID, &o.SpanID, &sample); err != nil {
+			return nil, fmt.Errorf("scan issue occurrence: %w", err)
+		}
+		o.Timestamp = ts
+		o.Sample = json.RawMessage(sample)
+		out = append(out, o)
+	}
+	if err := rows.Err(); err != nil {
+		return nil, fmt.Errorf("iterate issue occurrences: %w", err)
+	}
+	// The query is newest-first; replay oldest-first to preserve issue titles
+	// and trend buckets as they were originally observed.
+	for i, j := 0, len(out)-1; i < j; i, j = i+1, j-1 {
+		out[i], out[j] = out[j], out[i]
+	}
+	return out, nil
+}
+
 const maxSamplesPerIssueTrace = 20
 const maxUntraced = 20
 

internal/storage/store.go

@@ -134,6 +134,7 @@ type IssueTracesPage struct {
 // did this issue happen in" in one query (ClickHouseStore).
 type OccurrenceStore interface {
 	AppendOccurrence(ctx context.Context, occ Occurrence) error
+	RecentOccurrences(ctx context.Context, limit int) ([]Occurrence, error)
 	IssueTraces(ctx context.Context, issueID string, opts IssueTracesOptions) (IssueTracesPage, error)
 	IssueSources(ctx context.Context, issueID string) ([]IssueSource, error)
 }

k8s/deployment.yaml

@@ -264,6 +264,9 @@ rules:
 - apiGroups: [""]
   resources: ["pods", "events", "services"]
   verbs: ["get", "list", "watch"]
+- apiGroups: [""]
+  resources: ["pods/log"]
+  verbs: ["get"]
 - apiGroups: ["networking.k8s.io"]
   resources: ["ingresses"]
   verbs: ["get", "list", "watch"]